An Archeological Predictive Model for 
Fort Irwin, California: an approach to 
model development and technology transfer

An archeological predictive model at Fort Irwin, California, used the location of the existing sites and their relationship to the landscape features to characterize the landscape according to its potential to have sites. A key element of the methodological approach was that it easily is transferable to the cultural resource management staff and that it is understandable to people who do not have an extensive statistical background. The methods chosen center on the ArcGIS software (ESRI, Redlands, CA) already in use at Fort Irwin. The data about the sites and landscape are first analyzed for patterns using a simple logic that considers the relative density of sites across specific landscape attributes compared to the average site density. High values for site densities indicate higher suitability of a landscape feature. Suitability maps for each of the environmental characteristics are then overlaid to develop the final potential maps.

The output from the model was used successfully to guide the strategic placement of survey plots to optimize cultural resource investigations related to land acquisition activities in areas adjacent to Fort Irwin in 2002. From the results of those outcomes, it was clear that the efficacy of the model output is tied fundamentally to the characteristics of the input data. Of particular importance is the need for a single model to include sites that share common functions. In addition, the level of detail of the landscape variables and the exact combination of landscape factors for a given type of site was important. Unfortunately, the creation of a single site potential map is a lengthy process, hampering experimentation with various combinations of inputs and cross-disciplinary discussions of the output. The cultural resource staff was introduced to the GIS methods required to create site potential maps. Based on this experience and feedback from cultural resource staff, the development of a more automated method to create the site potential maps is currently underway. Software development for this method is guided by a need to read existing GIS files directly, to allow for better data exploration and graphing, and to reduce dramatically the time needed to create site potential maps.